.jpg)
Video analytics is one of the biggest breakthroughs in security technology in recent years, dramatically increasing the efficiency and effectiveness of video surveillance. Now, it's becoming more widespread than ever before. It is therefore incredibly important to have a clear understanding of this software may, or may not, be able to benefit your company.
THE BASICS: HOW DOES VA WORK AND WHAT ARE ITS CAPABILITIES?
Video analytics creates a complex mathematical model of the background scene of a camera’s view which enables objects in front of this background to be detected due to the pixel differences created. The software can therefore easily identify instances such as a breach of perimeter, someone entering a high risk zone, license plate recognition, or an object being removed or placed within the scope of vision. Once detected, the software produces an alert sent directly to security personnel for verification and investigation.
But how does the software know what should be identified as a suspicious event and what should be considered normal activity? Operators of the software can set specific rules depending on the risks at each site and the type of incident they want to detect. For example, rules can also be set for events such as crowd formation, loitering, or someone moving in the wrong direction of traffic, among many more. There are multiple different rules (often depending on which model in use) and more are continuously being developed. The software can therefore be tailored by security managers to target specific areas or scenarios they want to detect.
The alert sent on identification of an ‘unauthorized movement’ is most commonly in the form of an almost instantaneous video pop-up to the operator monitoring the site, but can also be sent to an email account or directly to the operator’s smartphone. Monitoring can be done from any location as long as the device has been set up within this system.
THE BENEFITS
Increasing efficiency and accuracy within your security system The key advantage of video analytics is that it is highly effective with accurate results which can result in significant cuts to the number of hours spent and the number of guards needed continuously manually monitor CCTV footage. Where humans are susceptible to lapses in concentration, video analytics operates 24/7, never ‘blinks an eye’ and can often detect incidents that security personnel may miss.
Moreover, VA enables operators to use CCTV footage to take immediate action during an incident and in some cases can prevent loss altogether through preempting crimes by early warning. Since security managers are alerted to suspicious activity in near real time, this helps to facilitate fast and informed decisions whereby operators can watch live footage of the scene.
VA can also provide major advantages in locating relevant CCTV footage after an incident has taken place. The software can perform query based searches to help end users quickly and efficiently find specific events, suspects or other required information, saving a significant amount of time that would be needed to locate such scenes manually.
Saving you money in the long run
The accurate analysis of footage during and after an event means that end users can consider significantly reducing the size of their security guard forces, therefore cutting the associated costs, making huge savings in the long run. In addition, another way that costs can be cut through the use of this technology is that operators can choose to only save recordings of suspicious events and simply discard the hours of insignificant footage with normal activity. This can save end users the costs involved in storing large amounts of useless footage.
Providing business intelligence
The advance functions that VA can perform has led to the software being used to collect data and trends from the footage recorded which can be used in business. Common uses that provide insightful data include people counting, customer patterns (including customer flow and areas of a store that have high and low levels of customer interest), traffic tracking and even queue monitoring. Access to this data can clearly have significant implications for the retail and commercial sectors. For example, understanding the flow of customers around a store, how long they spend queuing in peak times and what displays or products attract them most is likely to influence how a store manager organizes their products, manages their staff and customer service, and may make them rethink their store’s strategy altogether.
THE SHORTCOMINGS: THINGS TO LOOK OUT FOR
False alerts
Video analytics software is not perfect so there are instances when false positive detections are brought to the security manager’s attention. Better and more expensive models may reduce this chance, but it is difficult to eliminate this False Alert Rate (FAR) completely. Scenes with a constantly moving background, such as an escalator, a river or even busy areas with a high level of activity, increase the risk of a false alarm being detected. If the camera itself is moved, it may also have this effect. False alarms create additional work for the operator, who has to manually verify all alerts, and if there are a large number of false alarms, this can become irritating and frustrating. In extreme cases, a high FAR can result in implementation ineffectiveness.
Integration
Integrating some video analytics software into existing security systems can be difficult and costly so it is essential to understand what cameras and systems a specific model can support. There are a number of technological factors that can influence the performance of the software, from the resolution and frame rate of the cameras used, to the processing power and memory space of your central system. It is therefore highly recommended to consult an expert and have someone assess your existing technologies before investing in VA.
Environmental factors
It is important to consider carefully where cameras used in VA are placed. Both the angle of the camera and the distance to the object can influence several factors used in the software. The level of lighting is also very important: video analytics requires a minimum level of lighting to detect objects and function at all, but dramatic and abrupt changes in light may lead to false alarms. Changes in the weather, especially for outdoor cameras, may also influence the software’s ability to detect incidents or to minimize false alarms.
BEFORE MAKING A DECISION: KEY POINTS TO CONSIDER
Focus on the return on investment
Is this really the right application for your business or your site? Do you need a complex system with multiple functions, rules and the ability to detect a variety of different incidents, or can you save money buying a less complex model that still meets your needs? Although the costs installing video analytics can be high initially, you may be able to save on guard forces and their costs in the long run. Remember, it’s not necessary worth saving on buying cheaper VA software if its performance turns out to be dissatisfactory and creates additional work for operators due to a high FAR. Try to select a model that has a high Probability of Detection (POD) and a low FAR, they do exist.
Do your homework or consult an expert
Different video analytics models have different ways of working so you need to find the solution that best meets your application. For example, it’s important to ask questions such as ‘Do we want to save all our CCTV footage or only when suspicious activity is detected?’ or ‘Do we want to be able to watch live footage?’ Different answers to questions like this will effect what model you should select, and how it should be integrated into your system. One key pointer is to select a model that has color detection so that you don’t waste time searching the screen for what the software has detected as a suspicious event, it’ll be obvious.
0 comments:
Post a Comment